The choice of library to be used in a particular experiment is dependent upon many factors. Unfortunately, in the absence of detailed structural information about the target-ligand interaction, it is impossible to predict in advance which type of library will yield the most productive ligands. The Ph.D. 7 library consists of randomized linear 7-mer peptides, and may be most useful for targets requiring binding elements concentrated in a short stretch of amino acids. The Ph.D.-12 library consists of randomized linear 12-mer peptides, but with a diversity equivalent to the Ph.D.-7 library. It may be useful for targets requiring 7 or fewer defined residues for binding, but which cannot be contained within the 7-residue “window” of the Ph.D.-7 library. For example, the motif ASDXXXTXPY has only six defined positions, but cannot be present in the Ph.D.-7 library. Additionally, 12-mers are long enough to fold into short structural elements, which may be useful when panning against targets that require structured ligands. A caveat is that the increased length of the randomized segment may allow your target to select sequences with multiple weak binding contacts, instead of a few strong contacts. If, following three or four rounds of panning with a linear peptide library, there is no clear consensus ligand sequence, it is possible that the library simply does not contain any clones that bind tightly to the target, i.e., the most “fit” peptide in the library does not bind with high enough affinity to be selected. In this case a structurally constrained library such as the Ph.D.-C7C library might work better. Such libraries are especially useful for targets whose native ligands are in the context of a surface loop, such as antibodies with structural epitopes. Additionally, imposing structural constraints on the unbound ligand may result in a less unfavorable binding entropy, improving the overall free energy of binding compared to unconstrained ligands (O’Neil, K.T. et al 1992 Proteins 14, 509-515). A major disadvantage of structurally constrained libraries is that the constraint may “freeze out” a conformation required for target binding, preventing binding outright rather than improving affinity (McConnell, S. J. et al 1994 Gene 151, 115-118).